Our current knowledge about the rate of tibiofemoral joint space narrowing (JSN) in osteoarthritis (OA) of the knee comes from epidemiologic studies of the radiographic progression. All of these studies have used conventional techniques of knee radiography (i.e., standing anteroposterior (AP) views of the knee in full extension). The extended knee view, however, is inherently limited with respect to assuring the radioanatomic conditions associated with reproducible measurement of joint space width (JSW): alignment of the medial tibial plateau (MTP) with the x-ray beam and centering of the tibial spines beneath the femoral condyle. Several alternative protocols for knee radiography have been developed recently which use fluoroscopy to standardize knee flexion and rotation to achieve proper alignment and centering of the joint space in the x-ray image. Compare to the conventional extended view, these protocols have been shown to afford more precise measurement of JSW and to permit identification of OA knees with a more rapid rate of JSN. It is unclear, however, whether unstandardized positioning in previous studies using extended view radiographs has obscured our understanding of the nature and rate of progression of knee OA. The specific aims of this study are to obtain and analyze >600 paired radiologic knee examinations from 5 previous studies of the radiographic progression of knee OA (i.e., cohorts from Indianapolis, IN; Baltimore, MD; Bristol, UK, Nottingham, UK; Tecumseh, MI). The adequacy of MTP alignment and knee rotation in each image will be graded independently by 3 raters. Additional data will be recorded for each subject, including clinical characteristics (age, sex, race, height, weight) and evaluations of disease progression (i.e., Kellgren & Lawrence (K&L) grade, ratings of individual bony features of OA, JSW measurement for each image). The pairs of films will be reconstituted, and each pair will be categorized to reflect the degree of radioanatomic standardization for knee flexion and rotation (i.e., both films acceptable, one film acceptable, neither film acceptable). We will evaluate the effect of radioanatomic positioning by statistical comparison of groups of paired images representing varying levels of radioanatomic standardization with respect to clinical characteristics and observed OA progression (K&L grade, bony features, JSN). To the extent that unstandardized radioanatomic positioning has obscured past results, OA progression should be most uniform and rapid in the group of paired films representing the highest conformity to currently accepted standards for radioanatomic positioning of the knee.